Swiss AI For Public Good: A ProSocial AI Blueprint For The World?

Forbes

14-07-2025

Cow in Swiss Montains
Current Artificial intelligence development resembles a high-stakes race between tech giants. But Switzerland has chosen a different path. The Swiss AI Initiative's forthcoming large language model represents more than just another technological milestone, it embodies a transformative vision of how AI can be systematically designed, deployed and governed to serve humanity's highest aspirations.
The Swiss Approach: Public Infrastructure For Public Good
ETH Zurich and EPFL announced in July 2025 that they will release a groundbreaking large language model developed entirely on public infrastructure. Trained on the "Alps" supercomputer at the Swiss National Supercomputing Centre the new LLM marks a milestone in open-source AI and multilingual excellence.
This 70-billion parameter model, scheduled for release in late summer 2025, represents a radical departure from the proprietary, closed-source models that dominate today's AI landscape. The model represents a significant milestone in open AI development, offering multilingual fluency in over 1,000 languages and dialects.
The initiative emerged from the Swiss National AI Institute, created in October 2024 by ETH Zurich and EPFL to provide a long-term and national perspective on AI-based research, education and innovation. Rather than pursuing general-purpose models that compete with ChatGPT, Switzerland aims to develop large language models for specific applications in fields such as science, education, healthcare, robotics and climate studies.
The ProSocial AI Framework: 4 T's For Transformative Technology
The Swiss model exemplifies what scholars and practitioners call ProSocial AI — a comprehensive methodology to ensure artificial intelligence serves as a catalyst for human flourishing and planetary well-being. ProSocial AI offers principles for the design, deployment and governance of an approach that results in AI systems that are deliberately tailored, trained, tested and targeted to uplift people and planet. It represents more than mere ethical guidelines; it's a strategic reimagining of AI's role in society.
The ProSocial AI methodology centers on 4T's framework — Tailored, Trained, Tested, and Targeted — which are embodied by Switzerland's initiative. Let's examine how each principle manifests in this project:
Unlike generic AI models designed for broad consumer use, Switzerland's approach prioritizes specialized applications. The Swiss model is tailored to excel in domains where the country has established expertise and where AI can deliver maximum societal benefit. This targeted approach ensures that computational resources and research efforts focus on solving real-world problems rather than creating yet another general-purpose chatbot.
The multilingual capabilities spanning over 1,000 languages reflect Switzerland's commitment to inclusivity and global accessibility. By prioritizing linguistic diversity, the model ensures that AI benefits extend beyond dominant languages to serve marginalized communities worldwide.
The Swiss model's training on public infrastructure represents a fundamental shift toward transparency and accountability. Unlike proprietary models trained on undisclosed datasets, this initiative emphasizes open processes that can be scrutinized, validated and improved by the global research community.
This approach addresses critical concerns about AI bias, data quality and algorithmic transparency. By making the training process public, Switzerland enables other nations and institutions to learn from, replicate and improve upon their methodology.
ProSocial AI demands rigorous testing not just for technical performance but for societal impact. AI must be tailored to specific needs, trained on diverse data to reduce bias, tested for ethical performance and targeted at solving measurable societal challenges. Switzerland's focus on scientific, educational and healthcare applications ensures that the model undergoes domain-specific validation in areas where accuracy and reliability are paramount.
The public nature of the project enables independent researchers to conduct comprehensive evaluations, fostering a culture of continuous improvement and accountability that private initiatives often lack.
Rather than pursuing AI for its own sake, Switzerland's initiative targets specific societal challenges where AI can make a measurable difference. The focus on science, education, healthcare, robotics and climate studies reflects a strategic alignment with global priorities and Switzerland's national strengths.
This targeted approach ensures that AI development serves clear societal purposes rather than merely advancing technological capabilities. By prioritizing ethics, sustainability, and inclusivity, ProSocial AI ensures that AI becomes a forceful tool for fostering long-term health for individuals, communities, and the planet.
A Model For Global Collaboration
As of February 2025 Switzerland had no dedicated AI laws. Rather than a comprehensive approach to AI regulation, Switzerland is taking a tailored approach, focusing on implementing the Council of Europe's AI Convention and making targeted adjustments to existing laws. Instead of a Swiss equivalent of the EU AI Act, Switzerland is focusing on sector-specific adjustments and non-binding measures to promote responsible AI practices. By developing AI on public infrastructure and releasing it as open-source, Switzerland demonstrates that nations can maintain technological sovereignty while contributing to global knowledge commons. Their LLM initiative also signals a shift toward more collaborative and transparent AI development.
This model stands in stark contrast to the diverse regulatory approaches emerging globally.
The European Union has implemented the EU AI Act, which entered into force on 1 August 2024 and will be fully applicable on 2 August 2026. The EU's risk-based approach bans applications that create unacceptable risk, such as government-run social scoring of the type used in China, while subjecting high-risk applications to specific legal requirements. Meanwhile, China focuses on state control and AI ethics, while the US relies on a market-driven approach, lacking federal oversight but allowing state-level AI regulations. Singapore offers yet another model with the ambition that decisions made by AI should be explainable, transparent and fair. Their GenAI Framework is the first step towards fostering a trusted ecosystem for generative AI.
Switzerland's approach is particularly relevant as nations worldwide grapple with these varied AI governance challenges. The AI dilemma (balancing safety with innovation) is universal, yet each jurisdiction pursues unique solutions. The new model aligns with the EU framework while maintaining its commitment to innovation and openness, potentially offering a middle path that combines regulatory rigor with collaborative development.
Practical Takeaways: 4 A's For Global Implementation
ProSocial artificial intelligence is a transformative approach aligned with the United Nations Sustainable Development Goals. It advocates for AI systems centered on human values, equity, and global cooperation. And the Swiss AI Initiative offers valuable lessons for citizens and leaders worldwide in that pursuit. 4 A's can help citizens to start implementing ProSocial AI principles even if they do not live in Switzerland:
Examine the AI tools you use daily. Questions worth exploring: Are they transparent about their training data and decision-making processes? Do they reflect diverse perspectives and values? How much control do you have over the AI systems that affect your daily life, and what are the implications of relying on foreign, proprietary AI platforms?
Support organizations and initiatives that prioritize open-source AI development and transparent governance. Explore opportunities to advocate for AI literacy programs in your communities. Look for ways to support public AI infrastructure and research capabilities in your region, perhaps by backing initiatives that focus on specific domains where your community has expertise rather than competing directly with tech giants.
Explore possibilities to support policies that require transparency in AI systems that affect public welfare. Advocate for legal frameworks that incorporate the 4 T's approach for AI systems used in public services. Support international partnerships for sharing AI research and best practices, following Switzerland's example of contributing to global knowledge commons.
Recognize that AI development is a collective endeavor where public awareness and engagement can shape outcomes. Small actions — from choosing transparent AI tools to supporting open-source initiatives — contribute to a broader movement toward beneficial AI. Remember that the most sustainable AI strategies may emerge from collaboration rather than competition. Switzerland's approach suggests that communities can maintain technological sovereignty while contributing to shared global knowledge.
Switzerland's AI for Public Good initiative represents more than a technological achievement; it offers one possible roadmap for how communities might harness artificial intelligence to serve broader human aspirations. By exploring the ProSocial AI framework and the 4T's methodology, citizens worldwide can consider how AI development might serve not just economic interests but the broader goals of human flourishing and planetary well-being.
The Swiss model suggests that in the age of artificial intelligence, sustainable competitive advantage may come less from having the largest AI model than from having the most thoughtful, transparent, and socially beneficial one. As we navigate an AI-transformed world, Switzerland's approach offers one example of how technology might be developed as a force for good — if we choose to pursue that path.